This article was written by Bill James
This article was published in 1978 Baseball Research Journal
Since the invention of the hook, fishermen have entertained one another with tales of the big one that got away. Baseball men are not so lucky: they know how big the one was who got away. To the San Francisco Giants, the outfield that got away (Bobby Bonds, Garry Maddox, and George Foster) is no tall tale. It is a fact regrettably well established. But not, until now, well measured.
How do you tell what constitutes a good trade? While it is obvious, for example, that the Cardinals’ 1964 trade of Ernie Broglio for Lou Brock was a spectacular success, we have never been able to express that fact by any simple, succinct mathematical statement. How could we? To fill the void, I have developed a method of “scoring” a trade that attempts to reduce all of the contributions of the players exchanged to a single ratio.
Each player’s season was evaluated in accordance with 17 rules, 6 applying to pitchers and 11 to non-pitchers. For example, Lou Brock’s contributions to the 1964 Cardinals were evaluated at 12, 3 points for games played (103), 4 for batting average (.348), 3 for slugging average (.5 27), 1 for home run percentage (2.9), and I for stolen bases (33), reduced to 10 because Brock batted less than 500 times for the Cardinals. 10 points is an above-average total for a full season and an exceptional total for 103 games. Brock earned no points for defense, though they are available. Meanwhile, Ernie Broglio is credited with 2 points for his wins total (4), and 2 for other reasons, but when adjustments are made for his ERA and activity level, he winds up with a rating of 1 point contributed toward the Cubs’ success in the 1964 season.
A full explanation of how the rating-point system works would leave us no space for anything else, so we will move on and anyone interested in more details can contact the author. The system makes no claims to precision, but attempts to sort out levels of performance with a consistency which is convincing to the point of being, in a general way, undeniable. Take, for example, the 1972 National League. The MVP, Johnny Bench, was rated at “16”, the same rating given to Billy Williams (37 HR, 122 RBI, .333) and Cesar Cedeno (22, 82, .320 with 55 SB). Two other players, Joe Morgan (16, 73, .292, 58 SB) and Steve Carlton (27-10) rated even higher, at 17. Obviously there is some question about who should rate 16 and who 17, as Bench, at 16, won the big award. But there is no doubt whatsoever that, in naming Bench, Williams, Cedeno, Morgan, and Carlton, we have listed a group of players who rank (for that season) at the very top of the league’s stars.
Similarly, were we to take the players who ranked at “12” (Willie Stargell, Jose Cardenal, Jon Matlack) we would have a list of players who were, while still excellent ballplayers, undeniably not in the class of the 16’s and 17’s. The players listed at “8” include Burt Hooton (11-14), Joe Tone (.289), and Tim Foli (.241 with good defensive stats at SS), still undeniably not in a class (that year) with Stargell and Cardenal, but still clearly better than such players, rated at “4”, as Ernie McAnally (6-15), Bobby Valentine (.274 as a utility man) or Steve Garvey (.269 in half a season at third), who, in their turn, are still making a much larger contribution than such 0-value players as Jim York (0-1, 5.25 ERA) and Bob Fenwick (0,4, .180).
Armed with a system for evaluating seasons, the next task is to sort out the ground rules of a fair trade evaluation method. Question one is, “What constitutes a `trade’?” “Curt Motton for cash and two players to be named later,” for example-is that a trade or a sale? I decided not to rate any transactions involving cash, except for those listed as “sold in exchange for.”
Question two is, “What is a fair rating period?” One year is obviously not enough, but you can’t go back every season and update the records of trades made 10 or 15 years ago. You have to stop somewhere. I decided on 5 years as an acceptable period. There are trades for which this is misleading, of course, such as the George Foster trade, but as a general rule 5 years is quite enough for the direction of the result to be well established. There were many other ground rules needed, but only two other important enough to explain here. If a player were traded again within his five-year rating period, then the points he accumulated with his future team are credited to the team which acquired him in the original trade, as they will be counted against that team when the player is traded out of the system. If, on the other hand, the player is later sold, drafted, released, becomes a free agent, or is reassigned in a minor league transaction, then the count stops.
To see how these rules work, let us examine one trade in detail. In the winter of 1965, the Cincinnati Reds committed one of the most visible blunders in trading history when they exchanged their superstar outfielder, Frank Robinson, to the Baltimore Orioles for pitchers Jack Baldschun and Milt Pappas and outfielder Dick Simpson. Robinson’s contributions to the Orioles over the next five years were assessed at 16, 13, 8, 14, and 14 points, a total of 65. Simpson played for the Reds, occasionally, for two seasons, his accomplishments measured at 1 point in 1965 and 0 in 1966. From then on he was traded to four teams in two years, contributing two more points for a total estimated value of 3. Baldschun made no measurable contribution to the Reds in two seasons before returning to the minors, where he somehow became the property of the expansion Padres. His future major league success, such as it was, is irrelevant to the Reds and is not counted to them. Milt Pappas, the main bait for Robinson, pitched for the Reds for two full seasons, his value considered at 5 in 1966 and 8 in 1967, and part of another (0 value) before being traded to Atlanta, where he earned an additional 5 and 3 points in two seasons. In mid-season, 1970, he was sold to Chicago, and that stops the count. His total value: 21 points. The final score on that trade: Baltimore 65, Cincinnati 24. The final log on the Brock trade: St. Louis 54, Chicago 8.
To provide a basis for analysis, I rated every major league player for every season between 1963 and 1972, and every significant trade that I could find record of between the close of business, 1962, and June 15, 1973. Many players, of course, also had to be rated for the 1973-77 seasons. A few points from the complete study will serve to illustrate the general accuracy, and marginal inaccuracy, of the player rating system. First, the correlation between team total rating points and team wins is .92 (Pearson product/moment), indicating a very close relationship. Second, of the 20 MVPs in those years, 10 were rated as high as any player in the league, while 5 others missed by only one or two points. One MVP, Ken Boyer of the `64 Cardinals, won with a rating of 11, although Willie Mays rated 5 points higher at 16. However, if you added to the system a 2-point bonus for playing on a pennant winner and a 4-point bonus for leading the league in RBI, it would nominate 17 of the 20 MVP correctly. 14 of the 17 Cy Young winners and 16 of the 20 top rookies were also the highest-rated players available for the awards.
Finally, the results. What were the best trades of the period, (or, from the other side, the worst)? Number one on the lop-sided list was the October 1970 trade between Detroit and Washington. Bob Short, then owner of the Senators, had come to baseball following a largely successful run as owner of the Los Angeles Lakers, and his experience in basketball had led him to greatly overestimate what star value would mean to a baseball team. Denny McLain, for his part, had become a superstar by age 24, and it had led him to greatly underestimate the value of a reasonable training program. The results were enough to put an end to 70 years of baseball in the nation’s capital; Joe Coleman (45 pts.), Eddie Brinkman (36), and Aurelio Rodriguez (34) went to Detroit in exchange for Denny McLain (5), Elliot Maddox (9), and two zero-value players, a final count of Detroit 115, Washington 14-101 points given away in a single trade, the worst (or best) exchange of the period.
Close behind it is the 1971 trade in which Houston traded Cincinnati several championships, in the form of Joe Morgan, whose 86 points in the rated period are easily the highest of any player traded. The Reds also got Geronimo (40), Billingham (37), Menke (8), and Armbrister (1) for Lee May, Helms, and Stewart. Houston actually got a lot of talent out of the deal, as both May and Helms rated well for several years, but it doesn’t compare to the talent that they traded away. Final Score: Cincinnati 172, Houston 79.
Rating behind that are, in order, Gaylord Perry and Frank Duffy for Sam McDowell (92-5); Nolan Ryan and others for Jim Fregosi (83-4); Ken Singleton, Tim Foli, and Mike Jorgensen for Rusty Staub (126-48), Amos Otis plus for Joe Foy (83-6); and Randy Hundley and Bill Hands for Lindy McDaniel and Don Landrum (9 1-19).
That’s right-the Mets made three of those in two years. The Mets have, as their fans well know, traded themselves out of contention by giving away half of the American League’s All-Star team in fruitless attempts to find a third baseman or power hitter. But for consistency and durability in the Monty Hall give-away game, no one can begin to touch the astonishing series of “coups” pulled off by the San Francisco Giants. They made 25 rated trades in the 1963-72 period, almost all of them bad ones, for a sum total of 270-717, a deficit of 447 points, enough to win three pennants without help, and enough to turn seven last-place teams into pennant winners. That may seem extreme, but consider what would happen to a last-place team if you essentially gave them the best years of Hundley and Hands, Perry and Duffy, plus Matty Alou (his trade was 8-57), Orlando Cepeda (22-50), Stu Miller and John Orsino (10-51), Ron Hunt (0-27), George Foster and several others. The series of trades by which they ridded themselves of Maddox, Bonds, Ontiveros, Falcone, and Speier have not yet come up for review.
The best of traders? The Kansas City Royals, easily. Although they existed for only half of the rated period and made only 13 rated trades, five of them bad ones, they pulled off such “Godfather” bargains as the Amos Otis deal (83-6), the Lou Piniella for two non-prospects exchange (43-7), Cookie Rojas for a minor leaguer (39-0), the Freddie Patek steal (69-14), John Mayberry for Jim York (52-2), Hal McRae plus for Richie Scheiblum plus (48-6), and others at 32-3 and 10-0, for a net gain of 225 points. They have traded a bunch of journeymen into a championship team.
Since we measure the best by the bulk gain, not the ratio, the accomplishment is made more and not less impressive by the short span of time in which it was accomplished. But it is true that many of the other organizations have had “gain” periods when they traded well, but have forfeited the gains during other managements. Two other managements could challenge the Royals’ record. The Cubs during the Durocher years made the Hundley deal, plus Fergie Jenkins and Adolfo Phillips for two old pitchers (8 8-29) and Jim Hickman and Phil Regan for Ted Savage (5 5-2). The Orioles during the Harry Dalton years pulled off more heists than the Dalton Gang, topped by the Mike Cuellar caper (74-12) and the Robinson deal.
Having computed all this, the question now becomes, what can we do with it? Well, first of all, we could make it more accurate, and more complete. The rating system I have devised is a good one, but by no means perfect. Second, we could use it, I would hope, to evaluate trading strategies. Beyond saying that one should trade with San Francisco whenever possible, I really haven’t reviewed my own data with an eye to such strategic questions as, “What kind of risk are you really taking in trading an established player for a hot prospect?” and “Is it a solid strategy to try to trade for a star player coming off a bad year?” and “How safe is it to trade an everyday player for a pitcher?”
This method is simply a tool, and a rather crude one at that. I’m not saying we would learn anything by analyzing these questions, but we couldn’t get any more stupid. We have a possible method here of getting answers. Unlike ballplayers and managers, general managers have never had a “record” to contend with before. Now, the record book has one more chapter.